(19)
(11) EP 0 277 745 A2

(12) EUROPEAN PATENT APPLICATION

(43) Date of publication:
10.08.1988 Bulletin 1988/32

(21) Application number: 88300621.5

(22) Date of filing: 26.01.1988
(51) International Patent Classification (IPC)4G01N 24/08, G01N 24/12
(84) Designated Contracting States:
DE FR IT NL

(30) Priority: 27.01.1987 GB 8701695

(71) Applicant: BRITISH TECHNOLOGY GROUP LIMITED
London SE1 6BU (GB)

(72) Inventor:
  • Smith, John Alec Sydney
    Wimbledon London SW19 7QR (GB)

(74) Representative: Davis, Norman Norbridge et al
BRITISH TECHNOLOGY GROUP LTD 101 Newington Causeway
London SE1 6BU
London SE1 6BU (GB)


(56) References cited: : 
   
       


    (54) Methods and apparatus for detecting certain compounds


    (57) The detection of compounds containing both nuclei of a kind which exhibit nuclear magnetic resonance (nmr) and nuclei of a kind which exhibit nuclear quadrupole resonance (nqr), for example compounds containing both ¹H and ¹⁴H nuclei, is carried out by subjecting a substance suspected of containing such a compound to a magnetic field and repetitively to three parallel activities A,H and B. Activity A comprises two similar rf pulse sequences PA and PB at the nmr frequency. Activity H is a temporary change in the value of the magnetic field, which may be a reduction to zero. Activity B is the application of additional rf signals. Activity H and/or activity B are different during each of the respective pulse sequences PA and PB. The activity B rf signals are at one or more frequencies selected so that in combination with the activity H in at least one of the pulse sequences the subsequent nmr signal is modified if such a compound is present. The difference in magnitude of the nmr signals obtained following the two types of rf pulse sequence is then measured.




    Description


    [0001] The invention relates to methods and apparatus for detecting certain compounds. More particularly the invention relates to methods and apparatus for detecting compounds containing both nuclei of a first kind which exhibit nuclear magnetic resonance (nmr) and neclei of a second kind which exhibit nuclear quadrupole resonance (nqr). Examples of such compounds are a range of narcotics and explosives such as cocaine, heroin, TNT, RDX and PETN.

    [0002] Nuclei of the first kind usually have a spin quantum number of 1/2 and include the nuclei ¹H, ¹⁹F, ³¹P which give rise to strong and easily detectable nmr responses in the presence of a magnetic field. The nmr frequency of such nuclei is given by ω=γH where ω is the angular frequency, γ is the gyromagnetic ratio specific to that nucleus and H is the intensity of the magnetic field. The nmr frequency thus depends on the value of the applied magnetic field. Nuclei of the second kind have a spin quantum number greater than 1/2 and include ¹¹B, ¹⁴N, ²⁷Al, ³⁵Cl. They exhibit characteristic frequencies known as their quadrupole resonance frequencies. These frequencies are usually much less dependent on the intensity of the applied magnetic field. For a given nucleus theses frequencies depend on the molecular structure within which it is contained and therefore provide a sensitive identification of individual compounds containing such nuclei. The examples of narcotic and explosive substances quoted above are compounds containing both the ¹H nucleus and the ¹⁴N nucleus and the invention is particularly useful in the detection of compounds containing these nuclei, although the invention can also be used to detect compounds having defferent combinations of nuclei.

    [0003] It has been suggested that by causing any nqr response, if present, to change the intensity of an induced nmr signal by transfer of energy betwen the two kinds of nuclei the presence of substances containing both kinds of nuclei could be detected. This phenomenon is known as double resonance and can be made to occur, for example, when the nmr frequency is made to coincide with the nqr frequency, a condition known as level-crossing.

    [0004] One such suggestion is described in British Patent No. 2057135. In that specification it is proposed to generate an nmr echo signal by the application of a pair of rf pulses of appropriate frequency in a magnetic field. The magnitude of this echo signal is measured. The magnitude of the magnetic field is then changed to a value at which the nmr frequency coincides with the nqr frequency of a nucleus of the second kind in a compound which it is desired to detect so shortening the nmr spin-lattice relaxation time, a phenomenon known as cross-relaxation. The magnetic field intensity is then returned to its initial value and a similar pair of pulses applied and the echo signal again measured. If there has been cross relaxationwhen the intensity of the magnetic field was at its changed value then the echo signal response after the second pair of pulses will be different. This will indicate the presence of the compound. However, the two-pulse sequence described in intended to generate a so called "solid-echo" signal and to be effective the two pulses must be in phase quadrature.

    [0005] Another method is described in US Patent No. 4,296,378. However the method described there does not distinguish between substances in which cross-relaxation occurs and those substances containing nuclei of the first kind with similar nmr spin-lattice relaxation times.

    [0006] Another suggestion for detecting the presence of certain substances is described in published PCT Application No. WO84/04173. The method described there is based on the assumption that in substances of interest the ¹H nuclei have long spin-lattice relaxation times, (T₁) and short spin-spin relaxation times (T₂). Three are, however, many hazardous materials which it may be desired to detect which do not fit into this category. One example is ammonium nitrate in which T₁ is reatively short and T₂ is relatively long. Another example is cocaine where T₁ is short both in high and low magnetic fields.

    [0007] It is an object of the invention to provide a method and apparatus capable of detecting any one of a large range of compounds in which double resonance can be made to occur.

    [0008] According to the invention in one aspect a method of detecting the presence of one or more compounds containing both nuclei of a first kind which exhibit nmr and nuclei of a second kind which exhibit nqr comprises subjecting material suspected of containing such a compound to a magnetic field and successively and repetitively to two types of sequence, each sequence comprising an rf pulse train at the nmr frequency of the first kind of nucleus in the said magnetic field followed by a time interval in which in at least one of the sequences there is a predetermined change in the magnitude of the said magnetic field, and during the time interval in at least one of the types of sequence applying rf signals at one or more frequencies selected so that in combination with the aforesaid change in the magnitude of the magnetic field the subsequent nmr signal is modified if such a compound is present, and measuring the difference in the magnitude of the nmr signals obtained following the two types of sequence.

    [0009] Where rf pulse trains are employed then these trains can comprise spin locking pulse sequences comprising 90° pulses of the form 90° ₀-τ-(90°₉₀-2τ)n where τ is a time delay and n is an integer of value one or more. Alternatively the rf pulses may comprise single or multiple 90° pulses. The two types of sequence may be applied alternately. Alternatively, a plurality of one type of sequence is first applied and is then followed by an equal plurality of the other type of sequence.

    [0010] In one embodiment of the invention the changes in the magnitude of the magnetic field are changes to a zero or low value and rf signals at the nqr frequency of one or more compounds which it is desired to detect are applied in the time interval of one of the types of sequence. In addition rf signals not at the nqr frequency of any of the compounds of interest may be applied during the corresponding time interval of the other of the types of sequence.

    [0011] In another embodiment of the invention the change in the magnitude of the static magnetic field during one of the types of sequence is a stepped change with at least one of the steps being held at a value in which level crossing between nuclei of the two kinds takes place if corresponding ones of said compounds are present. The rf signals are applied when the magnetic field reaches a zero value in order to destroy any residual nmr signal.

    [0012] In yet another embodiment of the invention the changes in the magnitude of the static magnetic field are different in the resepective time intervals of each of the types of sequence. One of the changes is such as to cause the nmr frequency of the first kind of nuclei to coincide withone of the nqr frequencies of nuclei of the second kind in a compound which it is desired to detect. Simultaneously therewith rf signals are applied at a frequency corresponding to a different nqr frequency of such nuclei.

    [0013] According to the invention on another aspect apparatus for detecting the presence of one or more compounds containing both nuclei of a first kind which exhibit nmr and neclei of a second kind which exhibit nqr comprises:
    means for applying a magnetic field to material to be tested;
    means for applying rf pulses to such material of predeter­mined radio frequency, intensity, duration and phase at the nmr frequency of nuclei of the first kind;
    means for temporarily changing the magnitude of the said magnetic field in a predetermined manner;
    means for applying rf signals to such materials during the time of said temporary change of magnitude of the said magnetic field said signals being of one or more predetermined frequencies; and
    means for measuring the free induction signal from nuclei of the first kind.

    [0014] In carrying out the invention separate rf coil means may be provided to which the said rf pulses and the said rf signals respectively are applied. Two separate receiving channels may also be provided to receive different ones of free induction signals from nuclei of the first kind.

    [0015] In order that the invention may be more fully understood reference will now be made to the accompanying drawings in which:-

    Figure 1 shows a timing diagram of rf pulses, switched magnetic field and applied rf signals that can be used in a method embodying the invention,

    Figure 2 shows in detail one rf pulse train that may be used in the method of Figure 1,

    Figures 3,4 and 5 show alternative timing diagrams of pulses, switched magnetic fields and applied rf signals used in methods embodying the invention,

    Figure 6 shows apparatus embodying the invention in block diagrammatic form, and

    Figures 7,8a,8b,9 and 10 give results obtained by using the apparatus of Figure 6 in accordance with one or other of the techniques shown in Figures 1,3,4, and 5.



    [0016] Referring now to Figure 1, there are three parallel activities that take place, which are labelled A, H and B. The A activity comprises the application of two types of sequence labelled PA and PB. Both types of sequence include similar rf pulse trains indicated by the compsite pulse envelope PSL folowed by a time interval τʹ. The activity H relates to the application of a magnetic field which is normally held at a constant value HO but within the interval τʹ it is switched to a zero value for a time τʹp which is less than τʹ. There is a delay time τd between the end of interval τʹp and of the interval τʹ and a further delay time given by τʹ-τd-τʹp between the end of pulse train PSL and the commencement of interval τʹp. In the time interval τʹp an activity labelled B takes place. Activity B consists of the application of an rf signal as shown by the B line. This signal may have one or more different frequencies as will be described hereafter.

    [0017] The pulse envelope PSL itself comprises single or multiple 90° pulses or one or more pulse trains in the form of a multiple pulse spin-locking sequence which is shown in more detail in Figure 2. The spin-locking pulse sequence comprises a first 90° rf pulse followed by a train of further 90° pulses whose rf phase is in quadrature with the first pulse. The first pulse is written as 90°₀ and subsequent pulses as 90°₉₀. The time spacing between the first pulse and the commencement of the series of 90°₉₀ pulses is τ and the time spacing between each of the 90°₉₀ pulses of the series is 2τ. The complete pulse sequence can therefore be written as 90° ₀-τ-(90°₉₀-2τ)n where n is an integer of value one or more. Provided that the frequency of the rf pulses is chosen relative to the magnitude of the applied magnetic field H₀ so as to satisfy the relationship ω=γHO where ω is the angular frequency of the rf pulses and γ is the gyromagnetic ratio of nuclei of the first kind, then nuclear magnetic resonance will be induced and a free induction signal will be generate after each individual pulse of the sequence. This signal peaks during the time interval 2τ between successive 90° pulses and its peak value is measured and stored for later use. The entire pulse train that is shown in Figure 2 is designated by the pulse envelope PSL in Figure 1.

    [0018] Turning now back to Figure 1 attention is directed to the line marked B which shows the application of rf signals during the time that the magnetic field H has a zero value. In the time interval in the sequence PA this signal is chosen to have a frequency which is not equal to any of the nqr frequencies of compounds suspected of being included in the material under investigation. However, in the sequence PB one or more frequencies are applied either simultaneously or sequentially which are made equal to at least one of the nqr frequencies of suspected compounds.

    [0019] In use of the timing diagram shown in Figure 1 the sequences PA and PB are applied successively and repetitively. The actual overall sequence can for example comprise a series of PA sequences followed by an equal series of PB sequences, as PAPAPA...PBPBPB.... Alternatively the sequences can be applied alternately i.e. PAPBPAPB...

    [0020] Measurements of the free induction signals for both the PA and the PB sequences are taken using quadrature phase-sensitive detectors and are added separately for each sequence. The summed free induction signals may be Fourier transformed to obtain the absorption-mode signal which is then integrated to improve the signal-to-noise ratio. If no suspected material is present then there will be no difference in the magnitudes of the two integrations. Howevger, if any such material is present then there will be a difference. The existence of this difference indicates the presence of one or more suspected compounds containing nuclei at the frequencies of the B signals applied during pulse sequence PB.

    [0021] To facilitate understanding of the way in which the above timing diagram functions, the two kinds of nucleus that may be present in a material being examined are termed in the following discussions as A nuclei and B nuclei. The application of successive and repetitive sequences PA and PB may generate pseudo-equilibrium conditions in the A-spin system in the sense that the A populations are maintained away from their thermal equilibrium values. Use of pseudo-equilibrium conditions has the advantage that the material being examined can be immediately subject to the above described pulse sequences and there is no need to wait for the material to acquire a finite magnetisation in the applied magnetic field.

    [0022] The magnetic field as shown in the H line in Figure 1 is cycled to zero adiabatically but slowly enough to ensure that the population ratios of both the A and B spin systems become equal at level crossing during the change. Level crossing will occur as the magnetically-split A-spin levels, equal to γAH, collapse to zero (or increase from zero) and cross over the B-spin levels, the quadrupole interaction of which ensures a non-zero frequency, ωB, even in zero field. At level crossing, when the condition ωB =γAH is satisfied, the A and B spin temperatures TA,TB defined by the population ratios

    become equal in a time TCP, the cross-polarisation time. The field cycling must therefore be slow enough in this train to ensure that the time spent at the level-crossing condition be longer than TCP.

    [0023] During the time interval τpʹ in sequence PA, off-resonant rf radiation of angular frequency ω, with or without frequency modulation may be applied, as shown by the B line in Figure 1, to the B spins in zero or low magnetic field. Frequency ω is off-resonant with respect to the B quadrupole resonance frequencies ωQ. During time τpʹ in sequence PB the B-spins are subject, in zero or low magnetic field, to resonant rf radiation with or without frequency modulation at saturation level at one or more of their quadrupole resonance frequencies ωQ(¹), ωQ(2), at the temperature of the object. For a three-level spin system like ¹⁴N (I=1), with three resonant frequencies labelled ωXyZ, usually ωx and/or ωZ will be saturated and the power spectrum of the B pulse can contain any number of different ωx (and/or ωZ) frequencies characteristic of all nitrogen-containing compounds in the sample whose presence is to be detected. In this way all such compounds are simultaneously irradiated. On remagnetising adiabatically the saturated ¹⁴N (B) levels pass on their energy to the ¹H (A) levels on level crossing. The magnetic field and irradiation cycle may be repeated several times during the time τpʹ to produce multiple level-crossing and enhance sensitivity. Finally, the reduced ¹H signal is observed by a free induction signal accumulator in the time sequence PB relative to the corresponding signal in the time sequence PA.

    [0024] This timing sequence shown in Figure 1 is related to that used in double resonance by level crossing experiments (drlc). The technique is particularly suitable for materials in which both the A nuclei and the B nuclei have medium to long spin-lattice relaxation times T₁ and with a reasonably high proportion of B nuclei to A nuclei. This occurs for the A nucleus ¹H and the B nucleus ¹⁴N in many explosives.

    [0025] For compounds in which the cross-polarisation time TCP is long but the B nuclei are relatively abundant, as in the ¹⁴NO₂ group in some explosives, the technique of Figure 3 may be used. In this technique the pulse train PSL is followed by time intervals in which the magnetic field is changed as in Figure 1, but the nature of the change and the nature of the rf signals applied in the time intervals are different. For the highest sensitivity time τʹ should be equal to or longer than the time T₁ for the A nucleus. The magnetic field cycle is adiabatic but may be too fast to produce equalisation of the A and B spin temperatures during level crossing. During time τpʹ in sequence PA the magnetic field is cycled straight to zero and simultaneously the A dipolar system saturated with low frequency radiation, typically 10 kHz, to destroy any remaining A magnetisation. During time τpʹ in sequence PB the same sequence of events is used except that the field cycle is now conducted slowly enough for equalisation of the A and B spin temperatures at level crossing to occur e.g. by halting for a time TCP or longer at one or more level crossing fields in which the A nmr frequency matches in turn one or more of the B quadrupole resonance frequencies of each of the materials to be detected, both during demagnetisation and remagnetisation. During the latter the B nuclei quadrupole levels share their surplus populations, acquired during the former, with the A nuclei spins in sequence, producing an increase in the A signal. If T₁ (A) is short, time τʹ (and τd) may be made so short as to reduce or eliminate many A responses from solids usually encountered in the objects to be examined; only materials containing B nuclei whose quadrupole resonance frequencies are matched at the level crossing steps will then give significant signals.

    [0026] For compounds in which T₁(A)»T₁(B), τʹ can be made much shorter than T₁(A) and the magnetic field steps, now needed on remagnetisation only, are made to last much longer than the cross-polarisation time TCP so that considerable re-polarisation of the A spins by relaxation recovery occurs in the level-crossing field due to thermal contact with the B spins. Away from the level crossing condition re-polarisation is much slower. Again, fast repetition of the pulse units may reduce the background signal considerably.

    [0027] These pulse trains may be called double resonance by adiabatic remagentisation (dram).

    [0028] The technique of Figure 4 is appropriate for compounds in which the time T₁ for the A nucleus is much longer than T₁ for the B nucleus and the B nucleus is in low abundance, as for ¹⁴N in heroin and other narcotics. Under these circumstances, if the A Zeeman levels are matched in low magnetic fields with a pair of quadrupole-split B levels, after a short period TCP in which the two spin systems come to a common spin temperature, the combined system A+B thereafter relaxes with a spin-lattice relaxation time (T₁)eff which is the weighted mean of the two individual relaxation times T₁(A) and T₁(B)

    in which ε is equal to (or proportional to) the number ratio of B to A spins NB/NA. If ε T₁(A)/T₁(B)»1, the method has high sensitivity for the detection of B spins.

    [0029] In this sequence as shown in Figure 4, τʹ is set close to T₁ in the field H. During time τpʹ in sequence PA the magnetic field is cycled adiabatically to a magnetic field in which the A magnetic resonance frequency is close to but not equal to one of the B quadrupole resonance frequencies characteristics of the material it is wished to detect. During time τpʹ in sequence PB the magnetic field is cycled to a value such that the A magnetic resonance frequency is exactly matched to the peak B frequency response. The A-nucleus spin-lattice relaxation is much faster during sequence PB than during sequence PA, resulting in a faster loss of polarisation during PBthan P A and very different accumulated signals at the two detectors. This difference may be even further increased by level crossing effects in the appropriate circumstances e.g. when T₁(A) is short.

    [0030] In a simpler version of this cycle which is applicable when the measurement field H₀ and the cross-relaxation field H are close in magnitude, pulse evelope PSL comprises a saturation recovery sequence of 90° pulses with separations comparable to T₁(A) in field H₀. In sequnce PA, the magnetic field is set to H₀ and during PB to H for an equal time, the A nmr frequency being changed accordingly. The A-nucleus spin-lattice relaxation is much faster during the sequence PB than during sequence PA, resulting in a partial recovery of the A signal during sequence PB relative to that in PA and different accumulated signals at the two detectors.

    [0031] In another version of the cycle, the magnetic field during time τpʹ of sequence PB is set to half the magnetic field required for matching so as to observe the narrower two-proton relaxation jumps.

    [0032] In all versions of this cycle appropriate to spin-1 B nuclei such as ¹⁴N, rf signals are applied in time τpʹ in both sequences PA and PB of frequency equal to another quadrupole resonance frequency differing from that which is being cross-relaxed. If, for example, the cross-relaxation magnetic field is matched to one nqr frequency ωBʹ irradiation at another nqr frequency ωBʺ causes positive enhancement of the recovered A magnetisation, providing thereby both an increase in sensitivity and a highly sensitive identification of the material present in the object.

    [0033] This technique is related to a group of methods known collectively as double resonance by cross relaxation (drcr); it can detect ¹⁴N signals in crystalline solids, in plastics and also in liquids and gels with long correlation times (10⁻⁵ to 10⁻⁶s).

    [0034] The technique of Figure 5 is appropriate to compounds in which the quadrupolar nucleus B has half-entegral spin (eg I=³/₂) and T₁(B) is short such as the ³⁵Cl, ³⁷Cl nuclei in cocaine hydrochloride. During time τpʹ in sequence PA as shown in Figure 5 magnetic field is switched as fast as possible down to or close to zero and intense rf radiation applied at the B quadrupole resonance frequency (ωQ) of the compound to be detected at the given temperature. During time τpʹ in sequence PB the cycle is the same except that the rf radiation is off-resonance by a carefully specified shift Δ=ω-ωQ. During the second cycle the effective frequency "seen" by the B nuclei in the rotating frame is ω-ωQ=Δ, and if this lies within the A dipolar absorption edge the A spins become polarised provided that T₁(B) is short. Fast return to higher field with the minimum value of τd "captures" this polarisation within the A Zeeman levels resulting in an enhanced A signal if ω<ωQ and a diminished (or inverted) A signal if ω>ωQ. If the timing period τʹ is made so short that the A signal is saturated, any remaining signal in zero field being destroyed by an initial burst of radiation within the A dipolar absorption edge, the A detector will register no response as will the B detector unless the B-nucleus whose frequency is irradiated is present. As in the technique of Figure 3 it is possible in this way to eliminate or reduce much of the background A-signal from other A-containing materials in the object under examination.

    [0035] This technique is related to experiments based on double resonance by thermal mixing (drtm).

    [0036] Apparatus suitable for carrying out the methods shown in Figures 1,3,4 and 5 is shown in Figure 6. The apparatus has two rf channels. These channels provide the pulses and the signals for the A activity and the B activity respectively. The A rf channel comprises an A frequency source 1A, a pulse programmer and gate 2A, an rf power amplifier 3A and a coupling network 4A which couples the output of the rf amplifier 3A to an rf coil 5A. Likewise the B rf channel comprises a B frequency source 1B, a pulse programmer and gate 2B, an rf power amplifier 3B and a coupling network 4B which couples output of power amplifier 3B to an rf coil 5B.

    [0037] The two coils 5A and 5B are positioned to provide rf fields to at least part of material under investigation and in addition a dc magnet 6 of controllable intensity provides a static magnetic field in which the material is immersed. The intensity of the magnetic field is controlled by a magnetic field cycling control 7. Received rf signals are detected in coil 5A and the detected signals are fed to a two channel receiver 8 and thence to either one of two transient signal recorders 9A and 9B. Control of the frequency sources 1A and 1B, of the pulse programmers and gates 2A and 2B, the magnetic field cycling control 7 and the transient recorders 9A and 9B is carried out by means of a computer 10 which takes the signals recorded in the transient recorders 9A and 9B processes them and provides an output 11.

    [0038] The apparatus of Figure 6 is programmed by computer 10 to operate in accordance with any one of the methods shown in Figures 1, 3, 4 or 5. The transient recorders 9A and 9B are triggered to detect the whole free induction decay signal, or in the case of a spin-locking sequence, only at the echo maxima. In a solid these signals decay in a time which is closer to T than T₂, giving a much greater time for signal sampling. The sampling of the received signal can be performed for example by pulse gated integration or by fast digitisation and accumulation in the memory of the computer followed by Fourier transformation and integration of the absorption signal. It is desirable to use such accumulation techniques since the materials to be determined may be considerably diluted as, for example, being present in luggage and their signals will therefore be reduced by the presence of inert or harmless hydrogen-containing materials whose response is not changed when the selected perturbation is applied to the B spin system. In the case of a large object being examined, such a piece of luggage, it may be necessary to examine it section by section by means of a sequentially switched array of A and B rf coils in order to keep the power requirements for the two amplifiers 3A and 3B within reasonable limits.

    [0039] Examples of results obtained from the use of the apparatus of Figure 6 are described below with reference to Figures 7, 8, 9 and 10. In each case the most appropriate one of the methods of Figures 1, 3, 4 or 5 is used. The A nuclei in each case is the ¹H nucleus. The B nucleus is ¹⁴N as present in various narcotics and explosives or else the ³⁵Cl or ³⁷Cl nucleus present in cocaine hydrochloride.

    [0040] The technique of Figure 1 detects a change in the ¹H magnetisation due to heroin under the following conditions. In the A channel a single 90° pulse is used of pulse width 3 µs and the ¹H free induction decay at 40.18 MHz from a 3g sample is integrated in a box-car integrator of gate width 2µs and time constant 0.1s. The magnetic field cycle PAPA...PBPB... is set to τʹ=4s, τpʹ=0.6s, τd=0.1s and the B channel generates an rf field of amplitude about 0.2 mT. A change in single of about 35 mV is produced when the frequency in the B channel is set to 4.024 or 4.063 MHz at 295K. Figure 7 shows the complete high-frequency ¹⁴N spectrum produced by stepping the frequency after each sequence. The line is a doublet of splitting 39 kHz and has an overall width 59 kHz. Under similar conditions, cocaine free base ("crack") gives signals at 4.245 and 4.260 MHz of width 30 kHz at the same temperature. The invention is therefore capable of detecting and distinguishing between these two compounds at room temperature.

    [0041] The techniques of Figures 3 and 4 give strong ¹⁴NH₄+ and ¹⁴NO₃- signals from 3g samples of ammonium nitrate under the following conditions. The A-channel setting is as in the previous paragraph. The magentic field cycle is that described in Figure 4 with a single ¹H 90° pulse and τʹ=5.5s, τpʹ=0.3s and τd=0.7s, the field being stepped after each sequence. As shown in the spectrum of Figure 8a the NH₄+ signals are detected at level-crossing magnetic fields equivalent to ¹H frequencies of 48.9, 64.8 and 116.8 KHz. With the parameters changed to give τʹ=8.6s, τpʹ=1.2s and τd as before then the equivalent ¹H frequencies to detect the ¹⁴NO₃- ion are at 103.9, 449.1 and 549.0 kHz. This is shown in the spectrum of Figure 8b. A two-proton relaxation jump for the latter appears near 212 kHz and is sharp and strong enough to be used in detection. Mixtures of (NH₄)₂SO₄ and (NH₄)NO₃ show signals from both, whose relative intensity can be used to analyse for their relative proportions.

    [0042] The techniques of Figures 1 and 5 will detect signals from ³⁵Cl, ³⁷Cl in cocaine hydrochloride. The A channel settings are as before. The magnetic field cycle has τʹ=5s, τpʹ=0.6s, τd=0.1s and the B channel has rf radiation whose frequency is stepped from 1.7 to 2.8 MHz with an H₁rot field of about 0.15 mT. The ³⁵Cl signals are seen at 2.547 MHz and the signals from ³⁷Cl at 1.998 MHz at room temperature. The complete spectrum is shown in Figure 9. This ³⁵Cl frequency is unique to cocaine hydrochloride and is most unlikely to be found in any other materials encountered in luggage, parcels or the like.

    [0043] Figure 10 shows the vy lines near 3,400 KHz in the ¹⁴N spectrum of the high explosive RDX. The recording was made by the same method as was used for NH₄NO₃ but with timing settings of τʹ=50s, τpʹ=1s and τd=0.7s.


    Claims

    1. A method of detecting the presence of one or more compounds containing both nuclei of a first kind which exhibit nmr and nuclei of a second kind which exhibit nqr comprising subjecting material suspected of containing such a compound to a magnetic field and successive and repetitively to two types of sequence, each sequence comprising an rf pulse or pulse train at the nmr frequency of the first kind of nucleus in the said magnetic field followed by a time interval in which in at least one of the sequences there is a predetermined change in the magnitude of the aid magnetic field, and during the time interval in at least one of the types of sequence applying rf signals at one or more frequencies selected so that in combination with the aforesaid change in the magnitude of the magnetic field the subsequent nmr signal is modified if such a compound is present, and measuring the difference in the magnitude of the nmr signals obtained following the two types of sequence.
     
    2. The method as claimed in Claim 1 in which the rf pulse train comprises a sequence of 90° pulses.
     
    3. The method as claimed in Claim 2 in which the pulse train comprises a spin locking pulse sequence comprising 90° pulses of the form 90°₀-τ-(90°₉₀-2τ)n where τ is a time delay and n is an integer of value one or more.
     
    4. The method as claimed in Claim 1 in which the rf pulse is a single 90° pulse.
     
    5. The method as claimed in any one of the preceding claims in which the two parts of sequence are applied alternately.
     
    6. The method as claimed in any one of Claims 1 to 4 in which a plurality of one type of sequence is first applied and is then followed by an equal plurality of the other type of sequence.
     
    7. The method as claimed in any one of the preceding claims in which there are predetermined changes in the magnitude of the said magnetic fluid in both types of sequence.
     
    8. The method as claimed in any one of the preceding claims in which the changes in the magnitude of the magnetic field are changes to a zero or low value and rf signals at the nqr frequency of one or more compounds which it is desired to detect are applied in the time interval of one of the types of sequence.
     
    9. The method as claimed in Claim 8 in which rf signals which are not at the nqr frequency of any of the compounds of interest are applied during the corresponding time interval of the other of the types of sequence.
     
    10. The method as claimed in any one of claims 1 to 7 in which the change in the magnitude of the static magnetic field during one of the types of sequence is a stepped change with at least one of the steps being held at a value in which level crossing between nuclei of the two kinds takes place if corresponding ones of said compounds are present.
     
    11. The method as claims in Claim 10 in which the rf signals are applied when the magnetic field reaches zero value in order to destroy any residual nmr signal.
     
    12. The method as claimed in any one of Claims 1 to 7 in which the changes in the magnitude of the static magnetic field are different in the respective time intervals of each of the types of sequence.
     
    13 The method as claimed in Claim 12 in which one of the changes is such as to cause the nmr frequency of the first kind of nuclei to coincide with one of the nqr frequencies of nuclei of the second kind in a compound which it is desired to detect and simultaneously rf signals are applied at a frequency corresponding to a different nqr frequency of such nuclei.
     
    14. Apparatus for detecting the presence of one or more compounds containing both nuclei of a first kind which exhibit nmr and nuclei of a second kind which exhibit nqr comprising:
    means for applying a magnetic field to material to be tested
    means for applying rf pulses to such material of predeter­mined radio-frequency, intensity, duration and phase at the nmr frequency of nuclei of the first kind;
    means for temporarily changing the magnitude of the said magnetic field in a predetermined manner;
    means for applying rf signals to such materials during the time of said temporary change of magnitude of the said magnetic field said signals being of one or more predetermined frequencies; and
    means for measuring the free induction signal from nuclei of the first kind.
     
    15. Apparatus as claimed in Claim 14 in which separate rf coil means are provided to which the said rf pulses and the said rf signals respectively are applied.
     
    16. Apparatus as claimed in Claim 14 or Claim 15 in which two separate receiving channels are provided to receive different ones of free induction signals from nuclei of the first kind.
     




    Drawing